Hi, I have filed JIRA ticket SPARK-59114 <https://issues.apache.org/jira/browse/SPARK-59114> and will be opening a PR shortly. In my view its a serious issue during the caching of plan, though not appearing in a straightforward scenario
If it is deemed as valid bug, I wanted to share this tracking information to minimize any inadvertent duplication and ensure they are resolved using the original JIRA and PR ( which will be opened shortly. As mentioned in the bug detail, though the fix may be trivial ( or complex depending upon where and how it is most elegantly fixed), the deeper problem is the re-usable spark plans which can become completely in - compatible with the current SQL Conf , as it has happened in this case. Regards Asif
